Papers
2
Total Citations
27
H-Index
2
About
Jiaxi Sun is a rising researcher in computer vision and 3D scene understanding, with a focus on multimodal learning and neural representation. Their most impactful work, "MRFTrans: Multimodal Representation Fusion Transformer for monocular 3D semantic scene completion" (2024, 24 citations), introduces a novel transformer architecture that fuses RGB and depth information to reconstruct complete 3D semantic scenes from a single image—a critical step for autonomous navigation and robotics. This contribution addresses the long-standing challenge of inferring occluded geometry and semantics from limited sensory input. Sun also developed "C2Fi-NeRF: Coarse to fine inversion NeRF for 6D pose estimation" (2024, 3 citations), which leverages neural radiance fields to recover precise object poses from 2D images, offering a robust solution for augmented reality and manipulation tasks. Though early in their career, Sun’s work demonstrates a clear trajectory toward bridging representation learning and practical 3D perception, with their transformer-based fusion method already gaining traction in the community. Their research promises to advance how machines interpret and interact with complex, cluttered environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2C2Fi-NeRF: Coarse to fine inversion NeRF for 6D pose estimation3 citations · 2024